A Novel Method for Extracting Features of Partial Discharge Pulse Current Signal

Abstract

Partial discharge (PD) not only causes the aging of the internal insulation of the power transformer but also characterizes the deterioration of the insulation system. During the operation of the power transformer, different types of partial discharges (PDs) may occur inside. The opportune PD pattern recognition is of great significance to the maintenance and operation stability of the transformer. The pulse current method is a widely used PD detection method, and the feature extraction of pulse current signal is a key step in PD pattern recognition. The present feature extraction method of pulse current signal usually uses a complex algorithm and takes a long computing time, which is not suitable for field application. This paper proposed a fast and effective method for extracting features of PD pulse current signals. The features extracted by this method can correctly classify the three typical insulation defects in transformers.